\documentclass[fleqn,12pt]{article}

\usepackage{setspace,url,fullpage,latexsym,amsmath,amsthm,amssymb,natbib,graphicx,appendix,float,rotating,caption, subcaption,multirow,longtable,colortbl,graphics,graphicx,enumitem,pdflscape,epstopdf,verbatim}

\usepackage{lipsum}
\usepackage{filecontents}
\usepackage[dvipsnames]{xcolor}
\usepackage{hyperref}
\usepackage{cleveref}

\hypersetup{
	backref =       true,
	pagebackref  =  true,
	colorlinks =    true,
	linkcolor =     [rgb]{0,0.5,0.9},
	anchorcolor =   [rgb]{0.0,0.0,0.7},
	citecolor =     [rgb]{0,0.1,0.5},
	filecolor =     [rgb]{0.0,0.1,0.7},
	urlcolor =      [rgb]{0.0,0.0,0.7},
}
    

\setlength{\tabcolsep}{3pt}										


\doublespacing

\title{\large Political Investment Cycles in Democracies and Autocracies}

\author{Daehee Bak{\thanks{Texas Tech University. Email: {daehee.bak@ttu.edu}. An earlier version of this article was presented at the International Studies Association meetings, Atlanta GA, March 2016.  I thank James Hollyer, Hoon Lee, Chungshik Moon, and Stephen Nelson for helpful comments.}} \\  \vspace{.1cm}}

\date{}




\begin{document}


\maketitle
\thispagestyle{empty}
\setcounter{page}{0}

\bibliographystyle{apsr}

\begin{abstract}\singlespace{
\noindent
Extant research has shown burgeoning interest in whether host countries' political uncertainty is a significant impediment of foreign direct investment (FDI). Building upon the scholarly consensus on the adverse impact of political uncertainty on FDI, this article demonstrates that the extent to which investment climates are unpredictable varies cyclically, on the basis of election timing in democracies and leadership turnover in autocracies. Our empirical results show that in presidential democracies, FDI tends to slowly increase after an executive election, and then decline as the next executive election nears. However, we find that an electoral investment cycle is not found in parliamentary democracies where election timing is irregular, less predictable, and endogenous to domestic economic conditions. We also find that a resembling political investment cycle exists in autocracies not through electoral cycle but through leadership tenure cycle. The level of FDI inflows tends to be relatively low early in autocrats' tenure when political uncertainty is high, and rise as autocratic leadership tenure increases over time but eventually wane again as autocratic leadership is destabilized in the late period of power transition. Our findings indicate the existence of heterogeneous political investment cycles, depending on regime type.} 
\end{abstract}\newpage

% Long Abstract: Extant research has shown burgeoning interests in whether host countries' policy uncertainty is a significant impediment of foreign direct investment (FDI). Given the scholarly consensus on the adverse impact of policy uncertainty, this article demonstrates that the magnitude of policy uncertainty and unpredictability of investment climates varies cyclically, on the basis of election timing in democracies and leadership turnover in autocracies. We argue that political investment cycles driven by policy uncertainty are not homogeneous across different political systems. Our empirical results show that in (semi)presidential democracies, FDI tends to slowly increase after an executive election, and then declines as the next executive election comes near. However, we find that an electoral investment cycle is not found in parliamentary democracies where election timing is irregular, less predictable, and endogenous to domestic economic conditions. We also find that a resembling political investment cycle appears in autocracies not through electoral cycle but through leadership tenure cycle. The level of FDI inflows tends to be relatively low early in autocrats' tenure when policy uncertainty is high, and rise as autocratic leadership tenure increases over time but eventually wane again as autocratic leadership is destabilized in the late period of power transition. Our findings indicate the existence of heterogeneous political investment cycles depending on regime type.

\noindent

\subsection*{Introduction}

Over the past decade, scholars have scrutinized the political determinants of foreign direct investment (FDI), among which the adverse impact of political instability and unpredictability has attracted burgeoning research interest  \citep[e.g.,][]{Nigh85,EndersSandler96,Collier99,MurdochSandler02,Fielding03,MurdochSandler04,Brada06,BlombergMody05,Bussehefeker07,SulimanMollick09,Bussmann10}. This research program has identified numerous domestic political circumstances that are likely to discourage foreign investment. However, predicting the ebb and flow of FDI is a difficult task not only because not all circumstantial political conditions are easily observable to outsiders but also because the level of policy uncertainty is often private information held by the host government. 

Then, what would be the sources of information that are easily observable over time as well as indicative of the host country's political uncertainty? Among many factors, this article highlights the importance of a political cycle whose impact on FDI has yet to be thoroughly examined, particularly in autocracies. Specifically, this article attempts to answer the following questions. Do FDI inflows systematically vary over the course of the host country's political cycle? If so, what are the recognizable political turning points that shape an approximate political cycle in democracies and autocracies? 

In democracies, the most visible political turning point is an election, before and after which political uncertainty tends to be greater than at other times.  Thus, foreign investors are likely to perceive an election as a turning point that is exemplified by unpredictable policy changes and hence hesitate to make an investment decision until a new government's economic policy agenda becomes conspicuous.  Recent studies  that have examined electoral investment cycles for different types of investment, such as private fixed investment, corporate investment, and housing market performance,  show empirical evidence that \textit{political investment cycles} exist depending on election timing in democracies \citep{CanesPark12, JulioYook12, CanesPark14}. Although these studies reveal electoral investment cycles in democracies, the existence of \textit{electoral FDI cycles} is still far from obvious.

Furthermore, this article acknowledges that the impact of an electoral cycle on FDI should differ across different types of political institutions. For example, the policy implications of executive elections in presidential democracies should differ from those in parliamentary democracies. Election timing in parliamentary democracies is less regular and not totally exogenous to domestic political and economic conditions, thus making it difficult to predict a cyclical pattern of FDI inflows driven by institutionalized political uncertainty ex ante.  In autocracies, elections are not only less free and fair but also less institutionalized, so that election outcomes tend to be neither fully contested nor highly influential to the incumbent's leadership survival.  Consequently, it is difficult to juxtapose an electoral investment cycle in democracies with that in autocracies.  In this article, we examine whether an analogous political FDI cycle exists centered around leadership turnover in autocracies where a meaningful electoral turning point is lacking or absent.   
 
Our empirical results show that political investment cycles exist in both democracies and autocracies. Specifically, we find that FDI tends to decrease as an executive election nears in democracies, after which it slowly rises.  This inverted U-shaped electoral FDI cycle appears to be significant only in presidential democracies (including semi-presidential democracies) in which an executive election is often scheduled in a predictable and regular fashion. A resembling FDI cycle also appears to be significant in autocracies not through an electoral cycle, but through the leadership tenure cycle. The level of FDI inflows tends to be relatively low early in autocrats' tenure  (i.e., before power consolidation)  when political uncertainty is high, and rise as autocratic leadership tenure increases over time; however, it eventually wanes again as autocratic leadership is destabilized in the late period of power transition. 

FDI inflows do not simply cease around an election or autocratic leadership turnover, but are inclined to decrease. This finding indicates that FDI in times of high political uncertainty may be more tolerant to the risk of unexpected policy changes of the host country.\footnote{Some firms might adopt more aggressive and risk-taking investment strategies to preoccupy the markets of host countries.} However, our findings unambiguously reveal empirical evidence for the theoretical mechanism that suggests that foreign investors tend to adopt a ``\emph{wait-and-see}'' strategy particularly when political uncertainty is high. Generally, our findings indicate that domestic politics is a systematic driving force of the temporal variation in FDI inflows. In addition, our results add to the work on the political determinants of FDI by linking political cycle to FDI inflows, rather than by focusing on various circumstantial determinants. 

In the next section, we discuss the literature on the relationship between political uncertainty and FDI. Then, we relate the discussion to political cycles in democracies and autocracies. The next section presents the empirical methodology and main results, in addition to various robustness checks. 



\subsection*{Political Uncertainty and FDI}

Our theoretical framework is developed upon three premises that are succinctly articulated in \citep{DixitPindyck94}: (1) FDI is irreversible at least partially, (2) uncertainty is involved in the prospect of economic returns, and (3) investment timing is mostly flexible. Optimal investment decisions are contingent upon these three characteristics of FDI and, more importantly, we argue that they interact with the level of policy uncertainty shaped by host countries' domestic political environments. In this section, we first present theories of irreversible investment under uncertainty and investment timing.  

A large body of literature has examined how the political uncertainty of host countries affects the prospect of FDI inflows. Most notably, an early study by \citet{Rodrik91} noted private investors' tendency to withhold investment until the clouds of political uncertainty clear away, in particular showing that moderate amounts of \textit{policy uncertainty} counteract positive expectations on investment returns because policy uncertainty functions as an extra tax on investment, even when investors are treated as risk-neutral actors. Consequently, the adverse effect of political uncertainty is found to counterbalance the stimulative effect of policy reforms that can be reversed in the future. 

An important premise of this claim is that FDI is, at least partially, irreversible to varying degrees.\footnote{The theoretical framework of irreversible investment under uncertainty was introduced in detail in \citet{Pindyck91,Dixit92,DixitPindyck94}. Refer also to \citet{PindyckSolimano93} for the early literature on the theory of irreversible investment.} When making investment decisions, firms should be willing to bear sunk costs that may not be fully recovered. In addition, firms divest their investment and sell used assets, often reflecting unprofitable investment climates, with a considerable risk of losing sunk costs including transaction costs \citep{DixitPindyck94}. Furthermore, the host government may interrupt the exit process via regulations against sell-out. Whether FDI takes a form of joint-venture, M\&A, or greenfield investment, it is difficult to deny the irreversibility of FDI as well as the costs of reversing FDI. This point appears to be more straightforward when compared to foreign portfolio investment that is relatively easier to reverse. 

The importance of irreversibility looms large under political uncertainty. Firms have an incentive to delay their investment decisions, and wait to obtain more information on the host government's policy agenda and investment climates \citep[e.g.,][]{McDonaldSiegel86,Rodrik91,Caballero91,PindyckSolimano93,BertolaCaballero94}. When the credibility or sustainability of the host government's policies is questionable, the risk of policy reversal should be sufficiently high for foreign investors to hesitate to commit themselves to market entry. This mechanism provides investors with an incentive to adopt a “wait-and-see” strategy. In other words, the option value of waiting would be increasing in the level of host countries' political uncertainty. \citet{RivoliSalorio96} found that firms tend to postpone their investment decisions until policy uncertainty is significantly reduced. \citet{ChenFunke11} also highlights the adverse effect of institutional uncertainty on FDI, arguing that ``an uncertain environment is an essential factor in the flying geese pattern of FDI'' (608).\footnote{This delaying effect of policy uncertainty may well be mitigated when firms observe the sufficient benefits of early entry in times of severe competition \citep{RivoliSalorio96, Trigeorgis96}.}  

In sum, the literature suggests that we should expect that FDI inflows should vary with the magnitude of political uncertainty in an empirically meaningful way \textit{ceteris paribus}, which we believe is cyclical in both democracies and autocracies. In the next section, we discuss the temporal variations in the level of political uncertainty with respect to political cycles in democracies and autocracies.   


\subsection*{Political Investment Cycle in Democracies}

Since the seminal paper of \citet{Nordhaus75}, many scholars have examined the existence of a political business cycle (PBC) that refers to the idea that incumbents have an incentive to manipulate the economy through fiscal or monetary policy changes in order to obtain political gains \citep[e.g.,][]{Nordhaus75,Tufte78,RogoffSibert88,Keech95,Franzese02}. Subsequent studies revealed that the pattern of PBC should be examined in various contexts. For example, past studies suggest that the primary objectives of political parties are important determinants of different types of pre-election economic manipulation \citep{Hibbs77,Alesina87,Alexinaetal97,BloombergHess03}, and that the level of political security for the incumbent should significantly affect the extent of such an opportunistic behavior \citep{Schultz95,Krause05}. 

The underlying premise of these PBC theories is that election is a meaningful turning point that could engender opportunistic economic behavior. This intuition has also been applied to cyclical patterns of other economic indicators. For example, \citet{BernhardLeblang02} show that the expected return in excess for bearing risk (i.e., risk premia) exists more often during periods of political uncertainty in democracies such as an election  period because it is more difficult to accurately predict exchange rate movements. \citet{BlockVaaler04} reveal that economic agents in credit markets react to the incumbent's opportunistic behavior near election unfavorably, for example, by downgrading credit ratings in election years. 

From these studies, one can derive an important implication with regard to foreign investors' decision-making rationale. A main reason for politicians' opportunistic behavior in election years is that election outcome is not deterministic. Economic actors have an incentive to employ certain protective mechanisms against politicians' crafty policies driven by political motives.  Thus, elections are expected to exacerbate uncertainty over economic policy, thus increasing the unpredictability of investment climates. Specifically, investors tend to become more uncertain, particularly in the election period, regarding changes in taxation on foreign assets, industry regulations, and expropriation risks. 

More importantly, a new government may breach or renege on the initial terms of an FDI contract as they gain more bargaining leverage over time, as the obsolescing bargaining model suggests \citep{Vernon71}. In addition, leadership turnover may lead to different government preferences over different types of industry and trading partners.  For example,  if a new leader has maintained a close political relationship with a specific type of industry  (A) and had minimal political connections with another  (B), the new government is likely to provide more support and fewer regulations to A than B, and investors in industry  B should consider these political uncertainties over possible changes in government preferences.  In addition, in general, ``Firms are likely to view a possible shift in leadership from a market-friendly leader to a socialist leader as worse news than a possible shift in the other direction'' \citep{JulioYook12}.      

That is, foreign investors are highly sensitive to what type of monetary, tax, and regulatory policies a new government is likely to adopt. When MNCs gauge market entry to a country, election results and a new government’s preferences heavily affect their bargaining positions and long-term prospects of economic gains, because potentially dramatic economic policy changes are detrimental to foreign investments. Once investment capital is sunk, MNCs' obsolescing bargaining power makes firms more dependent upon government policy choices. Consequently, ``insofar as elections are predictable events, rational economic agents involved in contract negotiations are likely to time their activities to avoid the un- certainty revolving around elections'' \citep[][169]{GarfinkelGlazer94}. Examining electoral investment cycles for different types of investment other than  FDI, recent studies also provide empirical evidence that private and corporate investments tend to decline in the pre-election period \citep{CanesPark12, JulioYook12}.\footnote{Relying on a similar mechanism, \citet{CanesPark14} found, in their housing market analysis, that domestic economic activities decline before elections.}  

We could find some anecdotal evidence for our claim. According to the 2007 A.T. Kearney FDI Confidence Index survey, 23\% of respondents said that uncertainty surrounding 2008 U.S. presidential election is a critical factor that leads firms to maintain or decrease investments in the United States (A.T. Kearney 2007, 4). \nocite{Kearney07} For another example, a recent Turkish general election in June was a heated issue among investors, since Turkey has been an attractive investment choice for MNCs in recent years. Multiple news sources have shown, based on interviews and surveys, that investors tended to adopt `wait and see' approaches before and after the election. According to a Financial Times article, for example, ``... what investors dislike most of all is \textbf{\emph{uncertainty}} [emphasis added] -- and Turkey's new political picture looks set to offer only questions in the short to medium term. Foreign companies are likely to take a wait-and-see approach in the meantime'' \citep{fingar}.  

The Indonesian presidential election in 2014 also provides anecdotal evidence that foreign investors' investment decisions are heavily affected by political uncertainty around an executive election: 
\begin{quote}
\singlespacing
``... despite Indonesia's large pool of labour, relatively low costs and a growing middle class, many potential investors say they will \emph{\textbf{wait}} [emphasis added] until the election is decided. At the top of the list of foreigners with big money to spend is Taiwan's Foxconn Technology Group, the world's largest electronics contract manufacturer and one of the major suppliers to Apple Inc... Foxconn, listed as Hon Hai Precision Industry Co Ltd in Taiwan, is waiting for the new government to take office in October before deciding whether to go ahead with a \$1 billion manufacturing project in Indonesia, a company source had said'' \citep{DanubrataHung2014}.
\end{quote} 

Although we agree with the adverse effect of democratic elections on FDI inflows in general, the electoral implications on political uncertainty would have heterogeneous nuances across different types of political systems. Particularly when elections are institutionalized in a regular and competitive fashion such as executive elections in presidential democracies, electoral political uncertainty is expected to be substantial and cyclical. For example, the timing of an election is more foreseeable in presidential democracies than in parliamentary democracies where the government can call an early election. More importantly, election timing in parliamentary democracies is often endogenous to the economic conditions that are also interacting with FDI: e.g., several studies found that economic performance would be a significant determinant of calling an early election \citep{Ito90, CargillHutchinson91, Chowdhury93, PalmerWhitten00, Kayser06}. Thus, in parliamentary democracies, an election may not necessarily indicate a period with greater political uncertainty because of the incumbent government's incentives to utilize an early election to obtain large vote shares in times of good economy.\footnote{\citet{SmithAlastair04} finds empirical evidence that improving economic performance does induce electoral benefits in early elections.}  

In sum, although we expect an inverted U-shaped FDI pattern over an executive electoral cycle in presidential democracies, such an electoral investment cycle is expected to be less straightforward and observable in parliamentary democracies. From this discussion, we derive a hypothesis:

\textbf{Hypothesis 1:} \emph{In presidential democracies, FDI is expected to increase after an executive election, but decrease as the next election nears.}


\subsection*{Political Investment Cycle in Autocracies}

In this section, we discuss why the same argument about a democratic electoral investment cycle is unlikely to extend to autocracies. \citet{Alexander02} argues that political uncertainty in democracies can be characterized as institutionalized uncertainty as also noted by \citep{Przeworski1991}.\footnote{\citet{Alexander02} argues that democratic uncertainty should be interpreted in a technical sense rather than in a substantive sense because uncertainty over election outcome does not necessarily undermine the preexisting rules and norms of democratic governance.} However, in autocracies, elections are not only less institutionalized but also less competitive. The existence of elections in electoral autocracies does not necessarily mean that an election outcome is contested as fairly as in democracies. \citet{MagaloniKricheli10} show that ``many of the contemporary authoritarian elections are afflicted with hideous violence against voters who accuse the regime of electoral fraud'', and argue that authoritarian elections are often used for a symbolic purpose to signal regime stability (137).\footnote{Furthermore, they argue that ``Dictators interested in gaining access to international funds possess a strong interest in adopting multiparty elections because donors generously reward dictators who hold elections'' (135).} This different underlying rationale of holding elections implies that the mere existence of elections in autocracies is not a strong indication of significant policy changes in the near future (i.e., a greater level of political uncertainty). Furthermore, if an autocratic incumbent is secure in office regardless of election, the incumbent may have less incentives to manipulate the economy before the election, as also argued by \citet{Schultz95} and \citet{Krause05}. 

The lack of institutionalized leadership transition suggests that autocratic politics is temporally unpredictable rather than cyclical. However, we argue that an autocratic political cycle is shaped by autocrats' tenure in office. A commonly observed pattern in autocracies is that autocrats are exceptionally vulnerable to regime challenges early in their tenure, thus creating unpredictable political environments. Once they survive this early period of power contest, autocrats tend to enter the phase of power consolidation in which they successfully sort out core elite supporters and insulate themselves from unexpected regime disruptions. However, regardless of how long they remain in power, autocrats eventually encounter the challenges of new factions and enter the stage of power transition. Although this political cycle is unable to accurately describe a variety of idiosyncratic leadership changes, we contend that autocratic leadership tenure is one of the most reliable and observable indicators of an autocratic political cycle. \citet{Li09} also argues that ``the effect of leader tenure is more clear-cut in autocracy'' (1106), acknowledging the limitation of using electoral transition as an indicator of autocratic time horizon. In the following sections, we relate such a political cycle to the prospect of FDI inflows. 

Autocratic leaders are vulnerable to internal challenges particularly in the early period of power competition, whereas democratic leaders are less concerned with non-constitutional and/or violent leadership failure early in their tenure (or in the so-called honeymoon period). The first and foremost job, once autocrats enter the office, is to consolidate their dictatorial power base at any costs because of the lack of institutional protection mechanisms of their leadership because the early period of authoritarian leadership is often described as a period of power struggle \citep{Haber06, Geddes06}. \citet{Svolik09} also shows that autocrats are more likely to be removed by irregular means such as coup, particularly early in their tenure. Leadership instability in this early period assimilates electoral contest in democracies, which leads to  significant political uncertainty. Thus,  we expect that foreign investors should be highly uncertain  regarding the durability of authoritarian leadership and doubtful of policy sustainability, particularly early in autocrats' tenure.        

As autocrats survive the early period of power competition and their leadership solidifies, political uncertainty caused by leadership instability is ameliorated (for foreign investors). Under a solid authoritarian leadership, an election does not add much noise to the continuity of government policies. Therefore, with all else equal, we expect that FDI should increase as autocratic tenure increases. 

However, regardless of how long they remain in office, political uncertainty tends to increase again as autocratic tenure moves toward a late period of power transition, which is often marked by succession crises. \citet{Olson93} argues that ``most dictatorships are by their nature especially susceptible to succession crises and \emph{\textbf{uncertainty}} [emphasis added] about the future... the absence of any independent power to assure an orderly legal succession means that there is always substantial uncertainty about what will happen when the current autocrat is gone'' (572). Consequently, foreign investors are more likely to believe that, in this later period of power transition, the current government policy may discontinue in an unexpected manner. 

Furthermore, moving from the consolidation stage to late years of leadership vulnerability, autocrats and ruling elites are likely to become more predatory and more willing to develop political and economic barriers to protect their vested interests, which may have an adverse effect on the economic growth and crowd out new private investments. Thus, we expect that the increasing pattern of FDI should be reversed as autocratic tenure approaches the end. In sum, our discussion leads to the following hypothesis: 

\textbf{Hypothesis 2:} \emph{In autocracies, as tenure increases, FDI is expected to increase since a new leader enters the office, but eventually decrease later in the autocrat's tenure.}



\subsection*{Research Design}

\subsubsection*{Sample}
The unit of analysis is country-year, and we select (semi-)presidential democracies and autocracies using the regime coding by \citep{CGV10}. For robustness checks, we also test Hypothesis 1 including only presidential democracies. The temporal dimension is 1971 to 2008 for the democracy sample, and 1970 to 2004 for the autocracy sample.\footnote{The different temporal dimensions result from the data availability of leadership tenure from the Archigos database ($\sim$ 2004) for the autocracy sample \citep{Archigos}.} The number of autocratic countries covered in our analysis is 96, and the number of (semi-)presidential democracies is 59. 

\subsubsection*{Dependent Variable}
The dependent variable is \textsc{fdi} that is measured by FDI inflows in millions of US dollars. We use the natural logarithm of the measure to normalize the highly skewed distribution.\footnote{The formula used for the logarithmic transformation is $ln(FDI+1)$ if $FDI \geq 0$, and $-ln(1+ |FDI|)$ if $FDI < 0$.} The potential weakness of this FDI measure is the difficulty in making cross-country comparisons because the sheer amount of FDI inflows may not fully capture the relative importance of FDI in relation to the size of a country's economy. However, our theoretical focus is on the within-country temporal variations of FDI inflows over the course of political cycles rather than cross-country comparisons. Technically, the cyclical variations in FDI are estimated controlling for the size of economy and domestic market. 

Our main analysis uses this inflow measure rather than FDI as a share of GDP, which is another commonly used measure of FDI. However, we check the robustness of our results using the alternative measure (FDI as a percentage of GDP), and the results show that our main findings hold. The FDI data are drawn from the United Nations Conference on Trade and Development database \citep{UNCTAD2010}.

     
\subsubsection*{Independent Variable}

The independent variable for the (semi)presidential democracy models that test Hypothesis 1 is \textsc{election proximity} and $\textsc{election proximity}^2$. This variable is measured by the logged number of years since the last executive election. The higher the number, the greater proximity to the next executive election. The data on election timing are drawn from the National Elections across Democracy and Autocracy (NELDA) dataset \citep{HydeMarinov12}. Because Hypothesis 1 also suggests that the relationship between \textsc{fdi} and \textsc{election proximity} is inverted U-shaped, we include $\textsc{election proximity}^2$. 

Elections are more regular and predictable in presidential democracies than other regime types; however, the length of the election cycle significantly varies across countries.  Consequently, any empirical forecasting made from our models may not be equally applicable to all presidential democracies.   For example, three years since the last executive election do not necessarily mean that the next election is near or remains in the distant future. Thus, we create another independent variable that allows us to predict the nonlinear impact of election cycle that is comparable across different presidential democracies:   $\textsc{election quantile}$ and $\textsc{election quantile}^2$. This alternative measure normalizes the distance from the last executive election to the next one, ranging from zero to one. This quantile measure can be used mainly because an executive election is highly predictable in presidential democracies, particularly for foreign investors. If not, normalizing a political cycle is inappropriate: e.g., foreign investors are highly uncertain when an autocrat's tenure will end, hence making it unrealistic to find the relative location of autocratic tenure based on the uncertain ending point. Our primary findings hold using this alternative independent variable.       

The independent variable for the autocracy models that test Hypothesis 2 is \textsc{tenure} and $\textsc{tenure}^2$. Hypothesis 2 suggests that the influence of autocratic tenure is curvilinear with the coefficient estimate of $\textsc{tenure}^2$ being negative and significant, while not making a clear prediction about the approximate location of the reflection point. \textsc{tenure} is measured by the number of years in office for autocratic leaders, recorded by the Archigos database \citet{Archigos}. Multiple leaders are recorded in the years of leadership turnover, in which cases we select leaders who were in office on the first day of those years. We use the natural logarithm of this measure to capture the diminishing impact of tenure over time and to normalize the skewed distribution.

\subsubsection*{Control Variable}

We include several control variables that can explain the host country's political and institutional conditions. We include the level of \textsc{domestic instability} measured by the natural logarithm of the domestic conflict index taken from the Cross-National Time-Series (CNTS) Data Archive \citep{Banks13}.\footnote{This measure is a composite index that records the number of different types of mass and elite unrest such as purges, riots, government crises, assassinations, strikes, guerrilla wars, revolts, and anti-government demonstrations.} In addition, we include a dichotomous variable that records whether a country has experienced an \textsc{intrastate war} with at least 1,000 battle-related deaths using the UCDP armed conflict data \citep{Gleditsch02}.

We control for several economic indicators that capture a country's economic conditions and performance. We include \textsc{economic development} measured by the log of GDP per capita, \textsc{economic growth} measured by the annual percentage change of a country's real GDP, and \textsc{market size} measured by the log of total population \citep{HestonSummersAten11}. To capture a country's relations with other economies, we include \textsc{trade openness} measured by the sum of exports and imports divided by GDP \citep{HestonSummersAten11}, the \textsc{financial openness} measure created by \citep{ChinnIto08}, \textsc{bit} that measures whether a country has any bilateral investment treaties with OECD countries using the Bilateral Investment Treaties data \citep{UNCTAD2010}.\footnote{Although the literature reveals the positive impact of BITs on FDI in general \citep{EggerPfaffermayr04, NeumayerSpess05, ButheMilner08, Milner14, SalacuseSullivan05}, the empirical evidence is far from definite, as \citet{NewcombeParadell09} noted. Thus focusing on BITs with OECD countries is useful because OECD countries are largely responsible for global FDI outflows, and past studies found that BITs with OECD countries do have positive and significant effect on FDI \citep{NeumayerSpess05, TobinRose06}.} Finally, we include the natural log of \textsc{resource rents} that is measured by per capita value of oil and natural gas production \citep{Ross12}. All control variables are lagged by one year except economic growth. The summary statistics for all variables are presented in Table \ref{summary}. 

\begin{center}
[ Table \ref{summary} about here ]
\end{center}


\subsubsection*{Estimators}
Given the time-series cross-sectional data,\footnote{We examined whether the time series is stationary using augmented Dickey-Fuller (DF) tests \citep{DickeyFuller79,DickeyFuller81} and nonparametric Phillips and Perron (PP) tests \citep{PhillipsPerron88}, and found that the FDI time series in our sample is stationary.} we use the panel-corrected-standard-error (PCSE) with a first-order autocorrelation structure and country-specific fixed effects including country dummies, which has been commonly used in the previous FDI studies. However, as \citet{BeckKatz95} noted, the PCSE performance is poor when the panel data in the sample consist of a relatively small number of years while having a much larger number of cross-sections, such as in our sample. To ensure the consistency of our results against different estimators, we also use the Driscoll-Kraay (DK) estimator \citep{DriscollKraay98} with country-specific fixed effects, whose performance is not affected by the panel data structure. Additionally, this estimator is known to be consistent against spatial and temporal correlation in errors \citep{Hoechle07}. In addition, we check the robustness of our results using OLS with country fixed effects and a lagged dependent variable. 


\subsection*{Results}

We present the main empirical results in Table \ref{fdiinflow}. Models (1) and (2) test Hypothesis 1 for (semi-)presidential democracies. The results show that the relation between FDI and election proximity is inverted  U-shaped, as expected in Hypothesis 1. Conversely, FDI appears to increase to a certain reflection point over the executive election cycle, but thereafter decreases. This significant cyclical pattern of FDI also appears in the sample including only presidential democracies in Models (3) and (4) (thereby reducing the number of observations by approximately 40\%). This result holds in both PCSE and DK estimators. In Models (5) and (6), we find that autocratic tenure has an inverted  U-shaped relation with FDI as Hypothesis 2 suggests, and this relationship appears to be statistically significant. Again, this finding holds against different estimators.
\begin{center}
[ Table \ref{fdiinflow} about here ]
\end{center}

To observe a fuller picture of the curvilinear impact of different political cycles,  we run simulations using the estimates found in Table \ref{fdiinflow}, holding control variables constant at their mean or median values. These simulations help us to identify the reflection points after which FDI starts to decline, and to observe how uncertainty around the point estimates changes over the political cycles. 
\begin{center}
[ Figure \ref{fdiinflow_combine} about here ]
\end{center}

Figure \ref{fdiinflow_combine} presents the simulation results. First, we find that the level of FDI inflows appears to be greater in presidential democracies than that in autocracies. In the left graph, we find that FDI appears to increase to approximately 2.3 years since the last executive election. This result hints the average changing pattern of FDI across different presidential democracies. For example, in case of the United States, this reflection point indicates that more than half of the electoral cycle had passed. However, this average effect of electoral cycle would have different implications in presidential democracies with a relatively longer electoral cycle. Thus, we run the same regressions using election quantiles as the independent variables.  
\begin{center}
	[ Table \ref{quantile} about here ]
\end{center}

Table \ref{quantile} shows the results using \textsc{election quantile}. The results again comfort with our theoretical expectation presented in Hypothesis 1. \textsc{election quantile} has a significant inverted U-shaped relation with \textsc{FDI} regardless the different estimators. However, the reflection point should be interpreted quite differently. The simulation result in Figure \ref{fdiinflow_execquota} shows that FDI tends to increase up to 40\% of an election cycle, and then decreases over the remaining cycle as the next election becomes closer. In addition, we find that, on average, the level of FDI inflows is greater immediately after an executive election than before the next election, which indicates that foreign investors appear to observe greater political uncertainty before an executive election. However, the amount of FDI inflows remains lower in the very early period after an election than in non-election periods. In other words, foreign investors' confidence in policy continuity or sustainability appears to increase after election, but slowly dissipate as the next election approaches.
\begin{center}
	[ Figure \ref{fdiinflow_execquota} about here ]
\end{center}




In the right graph of Figure \ref{fdiinflow_combine}, we find that the level of FDI inflows is increasing to approximately 4.1 years in autocrats' tenure, after which FDI appears to decrease over the course of remaining tenure. In the autocracy sample used in the regressions, the average length of leadership tenure is approximately ten years. Thus, for an autocratic leader who remained in office for approximately ten years, FDI appears to increase to approximately 40\% of that autocrat's tenure. Substantively, the results predict that FDI is expected to increase by approximately twelve million dollars from the first to the second year in office, holding other variables constant at their mean or median values. The x-axis is not the number of years in office, but a logged measure of autocratic tenure; therefore, the decreasing pattern of FDI after the reflection point would be more gradual than it appears in the graph. 

       

The performance of control variables appears to be different between the autocracy and presidential-democracy samples. In both samples, \textsc{economic development}, \textsc{market size}, \textsc{bit}, and \textsc{financial openness} have significant and positive effects on FDI inflows. But \textsc{economic growth}'s positive effects appear to be significant only in the autocracy sample. Similarly, we find significant evidence that \textsc{intrastate war} and \textsc{domestic instability} have negative effects on FDI, only in the autocracy sample. These inconsistent results might be driven by the relatively small number of observations in the (semi)presidential-democracy samples.    


\subsection*{Robustness Checks}
In this section, we present the results from addtional empirical tests to verify the robustness of our empirical findings.\footnote{We thank the anonymous reviewers for suggesting many of these helpful robustness checks.} The results from these robustness checks are presented in detail in the Web appendix. 
 
 
 \subsubsection*{\emph{Institutional variations in presidential democracies}}
 
 Investors may be less concerned about political uncertainty in democracies where institutional checks and balance are well-established. That is, in mature democracies a leader may be more constrained in his or her ability to reverse the existing course of economic policies in an unexpected or idiosyncratic manner. Thus, Hypothesis 1 may have less empirical support in mature democracies than in nascent democracies. We test this conjecture by multiple means: controlling for democracy age, using split samples between mature and nascent democracies, and interacting regime maturity with electoral cycle. 
 
 The results in the appendix show that democracy age does not change our main findings substantively. However, we find minimal evidence that FDI inflows are affected by electoral cycle in mature presidential democracies, whereas the significant effect of electoral cycle holds in nascent democracies. The models that include the interaction term between regime maturity and electoral cycle also reveal that the significant inverted  U-shaped relationship between electoral cycle and FDI is not significant in mature democracies. In other words, our theoretical expectation in Hypothesis 1 holds solely for immature democracies, which indicates that electoral investment cycles are not homogenous even within presidential democracies.\footnote{We also examine whether the effect of electoral cycle is conditioned by a democratic country's economic conditions in the appendix:  e.g., the effect of electoral cycle may be smaller when a country is experiencing significant economic growth because a leader may have less incentives to reverse existing economic policies that have been proven to be successful. However, we find minimal empirical evidence for this conjecture.} 
 
 
 \subsubsection*{\emph{Disaggregating autocratic regimes}}
 
 We also examine whether political investment cycles differ across different autocratic regime types. For example, personalist or military dictators, who tend to be more myopic and predatory than those in single-party regimes, would not be able to generate much confidence in stable investment climates even when they have stayed in office for a long time. Although this conjecture does not necessarily counteract the overall autocratic investment cycle proposed in this article, we test these variations depending on autocratic regime type. Overall, the results in the appendix reveal that our theoretical prediction holds regardless of autocratic regime type. However, we also find that the amount of FDI inflows is greater in party regimes than in military/personalist regimes and that the inverted U-shaped FDI cycle is less straightforward in military/personalist regimes.   This nuanced finding indicates that autocratic investment cycle may also have significant variations depending on institutional features in autocracies.\footnote{In addition, we test if the empirical results for the autocracy sample are driven by major Asian economies under autocracy that have been popular destinations of FDI. The results in the appendix show that our findings hold excluding potential outliers such as Taiwan ($\sim$1995), China, South Korea ($\sim$1987), and Singapore.}   
 
 
 \subsubsection*{\emph{Developed vs. Developing Countries}}
 
 \citet{Blonigen05} demonstrate that pooling OECD and non-OECD countries in an empirical analysis is likely to lead to incorrect inferences. We rerun our main models splitting the samples between developing (non-OECD) and developed (OECD) countries. The results in the appendix show that our main findings hold only for developing countries, whereas the initial findings using the pooling sample do not hold in the OECD sample. Although this might be because FDI inflows are driven by different processes in developed and developing countries, we hesitate to read much into the findings from the OECD models because the number of observations and countries included in these OECD models is excessively small. 
 
 
\subsubsection*{\emph{Electoral investment cycle in parliamentary democracies}}

Our theory suggests that the proposed effect of election cycle on FDI should be blurred by the endogenous election timing in parliamentary democracies. We test this auxiliary theoretical conjecture, and find no statistically significant relation between legislative $\textsc{election proximity}$ (or legislative $\textsc{election quantile}$) and FDI inflows, while a similar curvilinear pattern appears. 


\subsubsection*{\emph{Electoral investment cycle in autocracies}}

Similarly, our theory hints that an election cycle should not have significant effect on FDI inflows in autocracies. The results in the appendix show that the empirical evidence for this claim is mixed. The coefficient estimate of $\textsc{election proximity}^2$ is \textit{positive} and significant, which indicates that FDI inflows tend to decrease after an autocratic executive election,\footnote{For example, the reflection point appears around 3.2 years after an executive election. The average length of an inter-executive election period is about 7.2 years in the autocratic sample.} but increase as the next election approaches. This finding further shows that the conventional notion of electoral investment cycle does not equally apply to autocracies.


\subsubsection*{\emph{Other robustness checks}}
We find empirical support for our hypotheses in general using FDI as a percentage of GDP. To address the argument that the preferences and ideological orientation of political parties are critical determinants of political business cycle, we confirm the robustness of our findings controlling for party information. In addition, we rerun our main models including year dummies to control for contemporaneous economic shocks, and the results in the appendix show that our main findings hold. Using OLS with a lagged dependent variable and country fixed effects yields similar findings. Furthermore, our findings hold controlling for political constraints on executive power, property righst protection, exchange rate, and economic crisis.  







\subsection*{Discussion}

Recent work on political investment cycle has revealed empirical evidence that private and corporate investment tends to be smaller when governments have greater political uncertainty driven by electoral cycle \citep{CanesPark12, JulioYook12, CanesPark14}. While this body of work provides preliminary evidence for the existence of political investment cycle, whether an analogous electoral FDI cycle exists has yet to be examined. This article fills this gap in the literature. Furthermore, there is a reason to believe that the patterns of political investment cycle are not homogeneous across different regime types. Our study adds to these extant studies by demonstrating that electoral investment cycles exist only in presidential democracies, and that political investment cycles in autocracies revolve around leadership turnover rather than election. Further, our empirical tests reveal the average turning points from which the temporal pattern of FDI inflows starts to change over democratic election cycles and during autocratic leadership cycles.       

Our findings appear to be robust against numerous robustness checks, while indicating that further examinations are needed to reveal more nuanced variations in political investment cycles within presidential democracies and autocracies. However, we should note several limitations of this study. The most important shortcoming originates from the potential counterargument to our theory; that is, there may be firms that have a strong incentive to preoccupy the domestic market of a host country, in spite of high political uncertainty due to an upcoming election or leadership transition. Conversely, the relatively less risk-averse sectors may be less resistant to the potential adverse effects of political uncertainty. This possibility implies that the shape and extent of political investment cycle may be different depending on FDI sectors.  Exploring this sector-dependent variation may be a meaningful extension of this study.

Another important caveat of this study is that autocratic leadership turnover is not always foreseeable for forward-looking foreign investors. We argued that autocratic leadership tenure is one of the most visible temporal indicators that can approximate the extent of political uncertainty, and circumstantial political and economic factors are not useful to reveal a cyclical pattern of FDI inflows. However, research using observational data cannot fully mimic the pre-investment risk assessment of investors, and neither researchers nor firms can accurately predict the timing of autocratic leadership transition. Although we control for some observable circumstantial factors that are believed to influence autocratic leadership turnover, such as economic growth, resource rents, internal warfare, and the level of domestic instability, we admit that our empirical models are not able to capture fundamental uncertainties in investors' expectations regarding the level of political uncertainty depending on autocratic tenure.\footnote{We further discuss an alternative measure, autocratic time horizon, that captures autocratic vulnerability in greater detail in the appendix.} 

In spite of this caveat, we believe that there are convincing reasons why the transitional period in authoritarian systems is plagued by political uncertainty, especially early in a new autocrat's tenure. This is quite contrasting to relatively secure leadership early in democratic leaders' tenure. Furthermore, not all domestic political and economic situations that are attributable to leadership transition are unobservable to foreign firms. Although our study does not fully explore the causal process given the limitation of observational data, the findings of this research still appear to be indicative of the existence of heterogeneous political investment cycles, depending on regime type.   





\clearpage
\setstretch{1.0}
\bibliography{fdicycle}





\clearpage
\setstretch{1.2}
\bgroup
\def\arraystretch{1.1}%
\begin{table}[H]
\begin{center}
     \caption{Summary Statistics} \label{summary}
     \scriptsize
     \begin{tabular}{lcccc}
     \hline\hline\\
				Variable & Mean  & Standard Deviation & Min & Max\\
			\hline			
			\hline			

			\emph{\textbf{(Semi)Presidential Democracies}} &&& \\
				ln(FDI Inflows) & 5.600 & 3.006 & -10.364 & 12.657 \\
				ln(Election Proximity) & 1.017 & 0.779 & 0 & 4.111 \\				
				Election Quantile & 0.508 & 0.364 & 0 & 1 \\								
				ln(GDP per capita) & 8.601 & 1.126 & 5.921 & 10.696 \\
				Economic Growth & 2.107 & 4.383 & -24.724 & 24.371 \\				
				ln(Population) & 9.398 & 1.272 & 6.020 & 12.617 \\
				Trade Openness & 59.018 & 30.421 & 9.876 & 175.959 \\
				BIT  & 0.699 & 0.459 & 0 & 1 \\
				Financial Openness & 0.168 & 1.478 & -1.875 & 2.422 \\
				Oil Rents & 2.376 & 2.505 & 0 & 8.528 \\
				Intrastate War  & 0.043 & 0.202 & 0 & 1 \\
				Domestic Unrest & 3.941 & 3.546 & 0 & 10.173 \\
			\hline
			\emph{\textbf{Autocracies}} &&& \\
			ln(FDI Inflows) & 3.187 & 2.952 & -7.070 & 11.013 \\
			ln(Tenure) & 1.946 & 0.983 & 0 & 3.850 \\				
			ln(GDP per capita) & 7.607 & 1.096 & 5.081 & 11.071 \\
			Economic Growth & 1.502 & 8.664 & -64.408 & 90.036 \\				
			ln(Population) & 9.902 & 1.375 & 6.176 & 14.066 \\
			Trade Openness & 64.566 & 44.357 & 6.616 & 373.983 \\
			BIT  & 0.690 & 0.463 & 0 & 1 \\
			Financial Openness & -0.540 & 1.282 & -1.875 & 2.422 \\
			Oil Rents & 2.161 & 2.872 & 0 & 9.616 \\
			Intrastate War  & 0.066 & 0.276 & 0 & 1 \\
			Domestic Unrest & 3.196 & 3.683 & 0 & 9.964 \\							

\hline\hline\\
   \multicolumn{5} {l} {\parbox[h][.2in][c]{4.5in}
{\scriptsize{\textbf{Note}: Descriptive statistics are calculated for the (semi)presidential-democracy sample (N=1,081) in Model (2) of Table 1 and for the autocracy sample (N=2,050) used in Model (6).}}}
\end{tabular}
\end{center}
\end{table}
\setstretch{1.0}





\clearpage
\setstretch{1.0}
\bgroup
\def\arraystretch{1.1}%
\begin{table}[ht]
\begin{center}
     \caption{Effect of Political Cycle on FDI Inflows} \label{fdiinflow}
     \scriptsize
     \begin{tabular}{lcccccc}
     \hline\hline
 \hspace{0.1cm} 				 & (1) & (2) & (3) & (4) & (5) & (6) \\
				 & (Semi)  & (Semi) &  &  &  &  \\
				 Sample:&  Presidential & Presidential & Presidential & Presidential & Autocracy & Autocracy \\
				Estimator: & PCSE & Driscoll-Kraay & PCSE & Driscoll-Kraay & PCSE & Driscoll-Kraay \\

\hline\\

				Election Proximity & 0.585** & 0.819*** & 0.645* & 1.015*** &  & \\
				& (0.241) & (0.238) & (0.359) & (0.347) &  & \\
				Election Proximity$^2$ & -0.357*** & -0.491*** & -0.361* & -0.572*** &  & \\
				& (0.133) & (0.144) & (0.210) & (0.210) &  & \\
				Tenure &  &  &  &  & 0.515*** & 0.606***\\
				&  &  &  &  & (0.184) & (0.201)\\
				Tenure$^2$ &  &  &  &  & -0.191*** & -0.217***\\
				&  &  &  &  & (0.060) & (0.055)\\
				Economic Development & 1.801*** & 1.550*** & 2.404*** & 2.069*** & 1.322*** & 1.147***\\
				& (0.510) & (0.428) & (0.731) & (0.456) & (0.320) & (0.300)\\
				Economic Growth & 0.003 & 0.021 & -0.015 & -0.009 & 0.023*** & 0.033***\\
				& (0.018) & (0.019) & (0.023) & (0.017) & (0.005) & (0.006)\\
				Market Size & 4.245*** & 4.138*** & 3.577*** & 3.639*** & 2.026*** & 1.775***\\
				& (0.718) & (0.521) & (0.920) & (0.685) & (0.402) & (0.460)\\
				Trade Openness & 0.010 & 0.000 & 0.016* & 0.010 & 0.002 & 0.003\\
				& (0.009) & (0.012) & (0.009) & (0.006) & (0.003) & (0.002)\\
				BIT & 0.566* & 0.732*** & 0.783** & 0.882*** & 0.907*** & 1.236***\\
				& (0.290) & (0.243) & (0.318) & (0.236) & (0.225) & (0.181)\\
				Financial Openness & 0.244*** & 0.284*** & 0.215* & 0.237*** & 0.309*** & 0.339***\\
				& (0.094) & (0.092) & (0.113) & (0.062) & (0.079) & (0.063)\\
				Resource Rents & -0.306 & -0.268** & -0.386** & -0.474*** & 0.157 & 0.211*\\
				& (0.209) & (0.106) & (0.181) & (0.138) & (0.110) & (0.113)\\
				Intrastate War & -0.049 & -0.301 & -0.357 & -0.552* & -0.393* & -0.638***\\
				& (0.285) & (0.309) & (0.278) & (0.273) & (0.223) & (0.199)\\
				Domestic Instability & -0.008 & -0.025 & 0.006 & 0.013 & -0.016 & -0.038***\\
				& (0.021) & (0.015) & (0.028) & (0.025) & (0.014) & (0.013)\\
				Constant & -60.238*** & -46.516*** & -57.860*** & -46.920*** & -31.127*** & -23.044***\\
				& (8.703) & (6.272) & (9.440) & (5.546) & (4.328) & (3.083)\\
				Observations & 1,081 & 1,081 & 656 & 656 & 2,050 & 2,050\\
				Number of Countries & 59 & 59 & 30 & 30 & 96 & 96\\
\hline\hline\\

\multicolumn{7} {l}
{\parbox[h][.17in][c]{5.6in} {\scriptsize{\textbf{Note}: All columns include country-specific fixed effects. (not reported). $^{*}$ p$<$0.10; $^{**}$ p$<$0.05; $^{***}$ p$<$0.01. (two-tailed tests)}}}
\end{tabular}
\end{center}
\end{table}
\setstretch{2.0}



\clearpage
\setstretch{1.0}
\bgroup
\def\arraystretch{1.1}%
\begin{table}[ht]
\begin{center}
     \caption{Effect of Election Quantile on FDI Inflows in Presidential Democracies} \label{quantile}
     \scriptsize
     \begin{tabular}{lcccc}
     \hline\hline
 \hspace{0.1cm} 				 & (1) & (2) & (3) & (4) \\
				  & (Semi) & (Semi) &  &  \\
				 Sample: & Presidential & Presidential & Presidential & Presidential \\
				Estimator: & PCSE & Driscoll-Kraay & PCSE & Driscoll-Kraay \\

\hline\\

\rowcolor[gray]{.95}				Election Quantile & 0.898** & 0.977* & 1.208** & 1.368**\\
				 & (0.440) & (0.484) & (0.584) & (0.665)\\
\rowcolor[gray]{.95}				Election Quantile$^2$ & -1.014** & -1.157** & -1.278** & -1.545**\\
				 & (0.425) & (0.461) & (0.568) & (0.668)\\
				Economic Development & 2.522*** & 2.319*** & 2.374*** & 2.034***\\
				 & (0.490) & (0.301) & (0.755) & (0.495)\\
				Economic Growth & 0.001 & 0.013 & -0.014 & -0.007\\
				 & (0.017) & (0.017) & (0.023) & (0.020)\\
				Market Size & 3.641*** & 3.580*** & 3.687*** & 3.874***\\
				 & (0.720) & (0.667) & (0.923) & (0.709)\\
				Trade Openness & 0.020*** & 0.020*** & 0.015 & 0.010*\\
				 & (0.007) & (0.004) & (0.009) & (0.006)\\
				BIT & 0.849*** & 1.018*** & 0.847*** & 1.002***\\
				 & (0.249) & (0.135) & (0.313) & (0.219)\\
				Financial Openness & 0.204** & 0.170*** & 0.224* & 0.217***\\
				 & (0.090) & (0.060) & (0.115) & (0.064)\\
				Resource Rents & -0.253** & -0.277*** & -0.371** & -0.445***\\
				 & (0.099) & (0.070) & (0.188) & (0.124)\\
				Intrastate War & -0.236 & -0.334 & -0.314 & -0.474*\\
				 & (0.275) & (0.272) & (0.282) & (0.278)\\
				Domestic Instability & -0.006 & -0.017 & 0.002 & 0.004\\
				 & (0.019) & (0.017) & (0.028) & (0.025)\\
				Constant & -60.667*** & -49.167*** & -58.966*** & -49.074***\\
				 & (8.352) & (5.202) & (9.525) & (5.835)\\ \\
				Observations & 1,018 & 1,018 & 631 & 631\\
				Number of Countries & 56 & 56 & 29 & 29\\
\hline\hline\\

\multicolumn{5} {l}
{\parbox[h][.16in][c]{4.2in} {\scriptsize{\textbf{Note}: All columns include country-specific fixed effects. (not reported). $^{*}$ p$<$0.10; $^{**}$ p$<$0.05; $^{***}$ p$<$0.01. (two-tailed tests)}}}
\end{tabular}
\end{center}
\end{table}
\setstretch{2.0}






\clearpage
\makeatletter
\setlength{\@fptop}{0pt}
\makeatother
\newpage
\begin{figure}
\begin{center}
\caption{Effect of Political Cycle on FDI Inflows}\label{fdiinflow_combine}  
{\includegraphics[width=6.0in,height=4.1in,trim= 0 0 0 0]{fdiinflow_combine.pdf}}
\end{center}
\vspace{-.25in}
\end{figure}



\clearpage
\makeatletter
\setlength{\@fptop}{0pt}
\makeatother
\newpage
\begin{figure}
\begin{center}
\caption{Effect of Election Quantile on FDI Inflows in Presidential Democracies}\label{fdiinflow_execquota}  
{\includegraphics[width=5.0in,height=4in,trim= 0 0 0 0]{fdiinflow_democracy_execquota.pdf}}
\end{center}
\vspace{-.25in}
\end{figure}






\end{document}
